Automated and Formal Synthesis of Neural Barrier Certificates for Dynamical Models

被引:25
|
作者
Peruffo, Andrea [1 ]
Ahmed, Daniele [2 ]
Abate, Alessandro [1 ]
机构
[1] Univ Oxford, Dept Comp Sci, Oxford, England
[2] Amazon Inc, London, England
来源
TOOLS AND ALGORITHMS FOR THE CONSTRUCTION AND ANALYSIS OF SYSTEMS, PT I, TACAS 2021 | 2021年 / 12651卷
关键词
HYBRID SYSTEMS; SAFETY VERIFICATION;
D O I
10.1007/978-3-030-72016-2_20
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
We introduce an automated, formal, counterexample-based approach to synthesise Barrier Certificates (BC) for the safety verification of continuous and hybrid dynamical models. The approach is underpinned by an inductive framework: this is structured as a sequential loop between a learner, which manipulates a candidate BC structured as a neural network, and a sound verifier, which either certifies the candidate's validity or generates counter-examples to further guide the learner. We compare the approach against state-of-the-art techniques, over polynomial and non-polynomial dynamical models: the outcomes show that we can synthesise sound BCs up to two orders of magnitude faster, with in particular a stark speedup on the verification engine (up to three orders less), whilst needing a far smaller data set (up to three orders less) for the learning part. Beyond improvements over the state of the art, we further challenge the new approach on a hybrid dynamical model and on larger-dimensional models, and showcase the numerical robustness of our algorithms and codebase.
引用
收藏
页码:370 / 388
页数:19
相关论文
共 50 条
  • [21] Automated adaptive inference of phenomenological dynamical models
    Bryan C. Daniels
    Ilya Nemenman
    Nature Communications, 6
  • [22] Automated adaptive inference of phenomenological dynamical models
    Daniels, Bryan C.
    Nemenman, Ilya
    NATURE COMMUNICATIONS, 2015, 6
  • [23] Neural closure models for dynamical systems
    Gupta, Abhinav
    Lermusiaux, Pierre F. J.
    PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES, 2021, 477 (2252):
  • [24] Synthesizing Barrier Certificates of Neural Network Controlled Continuous Systems via Approximations
    Sha, Meng
    Chen, Xin
    Ji, Yuzhe
    Zhao, Qingye
    Yang, Zhengfeng
    Lin, Wang
    Tang, Enyi
    Chen, Qiguang
    Li, Xuandong
    2021 58TH ACM/IEEE DESIGN AUTOMATION CONFERENCE (DAC), 2021, : 631 - 636
  • [25] Automated formal synthesis of Wallace Tree multipliers
    Hasan, Osman
    Kort, Skander
    2007 50TH MIDWEST SYMPOSIUM ON CIRCUITS AND SYSTEMS, VOLS 1-3, 2007, : 250 - 253
  • [26] FORMAL DESCRIPTION TECHNIQUES AND AUTOMATED PROTOCOL SYNTHESIS
    CARCHIOLO, V
    FARO, A
    GIORDANO, D
    INFORMATION AND SOFTWARE TECHNOLOGY, 1992, 34 (08) : 513 - 521
  • [27] A FORMAL TEXT DESCRIPTION OF MECHANISMS FOR THEIR AUTOMATED SYNTHESIS
    Dmitrochenko, Oleg
    Matikainen, Marko
    Mikkola, Aki
    PROCEEDINGS OF THE ASME INTERNATIONAL DESIGN ENGINEERING TECHNICAL CONFERENCES AND COMPUTERS AND INFORMATION IN ENGINEERING CONFERENCE, 2014, VOL 5A, 2014,
  • [28] Compositional Synthesis of Controllers via Co-Biichi Barrier Certificates
    Galarza-Jimenez, Felipe
    Murali, Vishnu
    Zamani, Majid
    IFAC PAPERSONLINE, 2024, 58 (11): : 57 - 62
  • [29] Automated Security Test Generation with Formal Threat Models
    Xu, Dianxiang
    Tu, Manghui
    Sanford, Michael
    Thomas, Lijo
    Woodraska, Daniel
    Xu, Weifeng
    IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, 2012, 9 (04) : 526 - 540
  • [30] Formal Synthesis of Lyapunov Neural Networks
    Abate, Alessandro
    Ahmed, Daniele
    Giacobbe, Mirco
    Peruffo, Andrea
    IEEE CONTROL SYSTEMS LETTERS, 2021, 5 (03): : 773 - 778